DOCSLIB.ORG

DRAFT1.2 Methods

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
DRAFT1.2 Methods HL7 v3.0 Data Types Specification - Version 0.9 Table of Contents Abstract . 1 1 Introduction . 2 1.1 Goals . 3 DRAFT1.2 Methods . 6 1.2.1 Analysis of Semantic Fields . 7 1.2.2 Form of Data Type Definitions . 10 1.2.3 Generalized Types . 11 1.2.4 Generic Types . 12 1.2.5 Collections . 15 1.2.6 The Meta Model . 18 1.2.7 Implicit Type Conversion . 22 1.2.8 Literals . 26 1.2.9 Instance Notation . 26 1.2.10 Typus typorum: Boolean . 28 1.2.11 Incomplete Information . 31 1.2.12 Update Semantics . 33 2 Text . 36 2.1 Introduction . 36 2.1.1 From Characters to Strings . 36 2.1.2 Display Properties . 37 2.1.3 Encoding of appearance . 37 2.1.4 From appearance of text to multimedial information . 39 2.1.5 Pulling the pieces together . 40 2.2 Character String . 40 2.2.1 The Unicode . 41 2.2.2 No Escape Sequences . 42 2.2.3 ITS Responsibilities . 42 2.2.4 HL7 Applications are "Black Boxes" . 43 2.2.5 No Penalty for Legacy Systems . 44 2.2.6 Unicode and XML . 47 2.3 Free Text . 47 2.3.1 Multimedia Enabled Free Text . 48 2.3.2 Binary Data . 55 2.3.3 Outstanding Issues . 57 3 Things, Concepts, and Qualities . 58 3.1 Overview of the Problem Space . 58 3.1.1 Concept vs. Instance . 58 3.1.2 Real World vs. Artificial Technical World . 59 3.1.3 Segmentation of the Semantic Field . 60 DRAFT version 1.0 22 Mar 1999 i 3.2 Technical Instances . 62 3.2.1 Technical Instance Identifier . 65 3.2.2 ISO Object Identifiers . 67 3.2.3 Technical Instance Locator . 71 3.2.4 Outstanding Issues . 72 3.3 Real World Instances . 73 3.3.1 Real World Instance Identifier . 74 DRAFT3.3.2 Postal and Residential Address . 84 Examples . 88 3.3.3 Person Name . 94 3.3.4 Organization Name . 114 3.4 Technical Concepts and the Code Value . 116 3.4.1 Outstanding Issues . 118 3.5 Real World Concepts . 120 3.5.1 The Concept Descriptor . 122 3.5.2 Code Translation . 123 3.5.3 Code Phrase . 124 3.5.4 Examples . 124 3.5.5 Outstanding Issues . 128 4 Quantities . 132 4.1 Overview . 132 4.2 Integer Number . 133 4.3 Floating Point Number . 134 4.4 Ratio . 137 4.5 Measurements . 138 4.5.1 Physical Quantities . 139 4.5.2 Monetary Quantities: Currencies . 140 4.5.3 Things as Pseudo Units . 143 4.6 Time . 144 4.6.1 Point in Time . 144 4.6.2 Time Durations . 147 4.6.3 Other issues and curiosities about Time . 147 4.6.4 Calendar Modulus Expressions . 148 5 Orthogonal Issues . 149 5.1 Interval . 149 5.2 General Annotations . 152 5.3 The Historical Dimension . 154 5.3.1 Generic Data Type for Information History . 154 5.3.2 Generic Data Type "History Item" . 155 5.4 Uncertainty of Information . 156 5.4.1 Uncertain Discrete Values . 158 5.4.2 Non-Parametric Probability Distribution . 159 5.4.3 Parametric Probability Distribution . 161 ii 22 Mar 1999 DRAFT version 1.0 5.4.4 Uncertain Value using Narrative Expressions of Confidence . 169 DRAFTAppendix A: All Data Types At a Glance . 171 DRAFT version 1.0 22 Mar 1999 iii DRAFT Abstract HL7 v3.0 Data Types Specification Version 0.9 Gunther Schadow DRAFTRegenstrief Institute for Health Care Abstract This document is a proposal for a complete redesigned set of data types to be used by HL7. Whereas in version 2.x data types where considered "formats" of character strings that would appear in HL7 data fields, this proposal assumes a more fundamental position: data types are the constituents of all meaning that can ever be communicated in messages. In HL7 v2.x, data types where defined a posteriori on an as-needed basis. Conversely this redesign defines data types a priori searching for fundamental semantic units in the space of all possible data types. This redesign work is heavily based on experiences with HL7 v2.x. Data types are defined for (1) character strings and multimedia enabled free text; (2) codes and identifiers for concepts and instances both of the real world and of technical artifacts; (3) all kinds of quantities including integer and floating point numbers, physical measurements with units, various kinds of time. Data types are classified (generalized) in various ways with respect to certain properties of interest. A number of issues have been identified to be equally applicable to many if not all data types. Intervals (of ordered types), uncertain information, incomplete information, update semantics, historic information, and general annotations are defined as generic data types, that can be used to enhance the meaning of any other type. Although this type system is precisely defined, it has a lot of flexibility not found in many other type systems. Precise conversions are defined between types so that data of one type can be used instead of another if there is a conversion. As a special case, character string literals are defined for most types which allows an instance of composite types to be sent in one compact character string. Copyright © 1999, Regenstrief Institute for Health Care. All rights reserved. DRAFT version 1.0 22 Mar 1999 1 HL7 v3.0 Data Types Specification - Version 0.9 Gunther Schadow 1 Introduction This document proposes a redesigned system of HL7 data types to be used for HL7 version 3. It is the result of a task force group spawned off Control Query at the San Diego Meeting in September 1998. Since then, that group has been meeting in weekly phone conferences, chaired by Gunther Schadow. The following people (mentioned in alphabetic order) contributed to this endeavor: James Case (University of California, Davis), Norman Daoust (Health Partners), DRAFTLaticia Fitzpatrick (Kaiser Permanente), Mike Henderson (Kaiser Permanente), Stan Huff (Intermountain Health Care), Matt Huges, Irma Jongeneel (HL7 The Netherlands), Anthony Julian (Mayo), Joann Larson (Kaiser Permanente), Randy Marbach (Kaiser Permanente), John Molina (SMS), Richard Ohlmann (HBO & Company), Larry Reis (Wizdom Systems), Dawid Rowed (HL7 Australia), Carlos Sanroman, Mark Shafarman (Oacis Healthcare Systems), Greg Thomas (Kaiser Permanente), Mark Tucker (Regenstrief Institute), Klaus Veil (Macquarie Health Corp., HL7 Australia), David Webber, and Robin Zimmerman (Kaiser Permanente). This task force planned to conclude its work by January 1999. Although we made tremendous progress due to the commitment of the task force members, we were not completely finished. By January (Orlando meeting) we were about 80% finished. By April 1999 (Toronto), we have about 90% of the work done. As usual, the last parts of a project consume the most amount of time and energy. However, all data types are defined by now and the remaining work is to polish and refine. This report is divided into two major parts. (1) The remainder of this introductory section explains the concepts and ideas that govern this proposed system of data types, while (2) the sections 2 through 5 will define the data types in detail. This document was compiled from the notes of the twentyfour (???) conferences. The conference notes where issued in Hypertext (HTML) and publicly available for browsing (http://aurora.rg.iupui.edu/v3dt). In the notes I heavily utilized the unique advantages of the hypertext medium, namely the ease by which one can follow cross references. It so happened that general concepts and detailed definitions were mixed together as they came up in the conferences. Hyperlinks have been an invaluable tool to recall definitions and explanations from earlier notes and to show how ideas evolved over time. This report is written as Hypertext too, but it is delivered to the general HL7 working group as a paper document, which required to bring the material into a systematic order. However, the division into a first part, explaining the overall concepts, and a second part, defining the data types in detail, is problematic, since the usefulness of the general concepts are illustrated only by how those concepts are actually used in the definitions of the data types. The definitions of the data types, however, depend on general rules. Thus the reader faces a kind of "hermeneutic circle", where one has to know about the first part before one can fully comprehend the second part and vice versa. The Hypertext version of this report contains numerous forward and backward links, which, in the printed form appear as cross references to page numbers in square 2 22 Mar 1999 DRAFT version 1.0 1.1 Goals brackets. This ordering of the material comes in handy for the "impatient reader" who can explore everything just by following cross references. The reader who wants to see just some actual type definitions can use the index [p. 171] and directly proceed to the types he or she is interested in. The reader who wants to read through all the data type definitions can directly proceed to the sections of the second part [p. 36] and, if necessary, follow links back to the explanation of DRAFTgeneral concepts. Those who want to read through all of the text from the beginning can start with the general concepts and will be guided forward to the points where each concept is actually used. A final word of acknowledgment. Many of the great ideas reported here are born in numerous and intense discussions that Mark Tucker and I had before and after the conference calls. Without Mark Tucker, this whole type system work would have never evolved to a useful state.
Load more

Recommended publications
  • A Hybrid Static Type Inference Framework with Neural
    HiTyper: A Hybrid Static Type Inference Framework with Neural Prediction Yun Peng Zongjie Li Cuiyun Gao∗ The Chinese University of Hong Kong Harbin Institute of Technology Harbin Institute of Technology Hong Kong, China Shenzhen, China Shenzhen, China [email protected] [email protected] [email protected] Bowei Gao David Lo Michael Lyu Harbin Institute of Technology Singapore Management University The Chinese University of Hong Kong Shenzhen, China Singapore Hong Kong, China [email protected] [email protected] [email protected] ABSTRACT also supports type annotations in the Python Enhancement Pro- Type inference for dynamic programming languages is an impor- posals (PEP) [21, 22, 39, 43]. tant yet challenging task. By leveraging the natural language in- Type prediction is a popular task performed by most attempts. formation of existing human annotations, deep neural networks Traditional static type inference techniques [4, 9, 14, 17, 36] and outperform other traditional techniques and become the state-of- type inference tools such as Pytype [34], Pysonar2 [33], and Pyre the-art (SOTA) in this task. However, they are facing some new Infer [31] can predict sound results for the variables with enough challenges, such as fixed type set, type drift, type correctness, and static constraints, e.g., a = 1, but are unable to handle the vari- composite type prediction. ables with few static constraints, e.g. most function arguments. To mitigate the challenges, in this paper, we propose a hybrid On the other hand, dynamic type inference techniques [3, 37] and type inference framework named HiTyper, which integrates static type checkers simulate the workflow of functions and solve types inference into deep learning (DL) models for more accurate type according to input cases and typing rules.
    [Show full text]
  • 5. Data Types
    IEEE FOR THE FUNCTIONAL VERIFICATION LANGUAGE e Std 1647-2011 5. Data types The e language has a number of predefined data types, including the integer and Boolean scalar types common to most programming languages. In addition, new scalar data types (enumerated types) that are appropriate for programming, modeling hardware, and interfacing with hardware simulators can be created. The e language also provides a powerful mechanism for defining OO hierarchical data structures (structs) and ordered collections of elements of the same type (lists). The following subclauses provide a basic explanation of e data types. 5.1 e data types Most e expressions have an explicit data type, as follows: — Scalar types — Scalar subtypes — Enumerated scalar types — Casting of enumerated types in comparisons — Struct types — Struct subtypes — Referencing fields in when constructs — List types — The set type — The string type — The real type — The external_pointer type — The “untyped” pseudo type Certain expressions, such as HDL objects, have no explicit data type. See 5.2 for information on how these expressions are handled. 5.1.1 Scalar types Scalar types in e are one of the following: numeric, Boolean, or enumerated. Table 17 shows the predefined numeric and Boolean types. Both signed and unsigned integers can be of any size and, thus, of any range. See 5.1.2 for information on how to specify the size and range of a scalar field or variable explicitly. See also Clause 4. 5.1.2 Scalar subtypes A scalar subtype can be named and created by using a scalar modifier to specify the range or bit width of a scalar type.
    [Show full text]
  • ACDT: Architected Composite Data Types Trading-In Unfettered Data Access for Improved Execution
    ACDT: Architected Composite Data Types Trading-in Unfettered Data Access for Improved Execution Andres Marquez∗, Joseph Manzano∗, Shuaiwen Leon Song∗, Benoˆıt Meistery Sunil Shresthaz, Thomas St. Johnz and Guang Gaoz ∗Pacific Northwest National Laboratory fandres.marquez,joseph.manzano,[email protected] yReservoir Labs [email protected] zUniversity of Delaware fshrestha,stjohn,[email protected] Abstract— reduction approaches associated with improved data locality, obtained through optimized data and computation distribution. With Exascale performance and its challenges in mind, one ubiquitous concern among architects is energy efficiency. In the SW-stack we foresee the runtime system to have a Petascale systems projected to Exascale systems are unsustainable particular important role to contribute to the solution of the at current power consumption rates. One major contributor power challenge. It is here where the massive concurrency is to system-wide power consumption is the number of memory managed and where judicious data layouts [11] and data move- operations leading to data movement and management techniques ments are orchestrated. With that in mind, we set to investigate applied by the runtime system. To address this problem, we on how to improve efficiency of a massively multithreaded present the concept of the Architected Composite Data Types adaptive runtime system in managing and moving data, and the (ACDT) framework. The framework is made aware of data trade-offs an improved data management efficiency requires. composites, assigning them a specific layout, transformations Specifically, in the context of run-time system (RTS), we and operators. Data manipulation overhead is amortized over a explore the power efficiency potential that data compression larger number of elements and program performance and power efficiency can be significantly improved.
    [Show full text]
  • Occam 2.1 Reference Manual
    occam 2.1 reference manual SGS-THOMSON Microelectronics Limited May 12, 1995 iv occam 2.1 REFERENCE MANUAL SGS-THOMSON Microelectronics Limited First published 1988 by Prentice Hall International (UK) Ltd as the occam 2 Reference Manual. SGS-THOMSON Microelectronics Limited 1995. SGS-THOMSON Microelectronics reserves the right to make changes in specifications at any time and without notice. The information furnished by SGS-THOMSON Microelectronics in this publication is believed to be accurate, but no responsibility is assumed for its use, nor for any infringement of patents or other rights of third parties resulting from its use. No licence is granted under any patents, trademarks or other rights of SGS-THOMSON Microelectronics. The INMOS logo, INMOS, IMS and occam are registered trademarks of SGS-THOMSON Microelectronics Limited. Document number: 72 occ 45 03 All rights reserved. No part of this publication may be reproduced, stored in a retrival system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior permission, in writing, from SGS-THOMSON Microelectronics Limited. Contents Contents v Contents overview ix Preface xi Introduction 1 Syntax and program format 3 1 Primitive processes 5 1.1 Assignment 5 1.2 Communication 6 1.3 SKIP and STOP 7 2 Constructed processes 9 2.1 Sequence 9 2.2 Conditional 11 2.3 Selection 13 2.4 WHILE loop 14 2.5 Parallel 15 2.6 Alternation 19 2.7 Processes 24 3 Data types 25 3.1 Primitive data types 25 3.2 Named data types 26 3.3 Literals
    [Show full text]
  • Floating-Point Exception Handing
    ISO/IEC JTC1/SC22/WG5 N1372 Working draft of ISO/IEC TR 15580, second edition Information technology ± Programming languages ± Fortran ± Floating-point exception handing This page to be supplied by ISO. No changes from first edition, except for for mechanical things such as dates. i ISO/IEC TR 15580 : 1998(E) Draft second edition, 13th August 1999 ISO/IEC Foreword To be supplied by ISO. No changes from first edition, except for for mechanical things such as dates. ii ISO/IEC Draft second edition, 13th August 1999 ISO/IEC TR 15580: 1998(E) Introduction Exception handling is required for the development of robust and efficient numerical software. In particular, it is necessary in order to be able to write portable scientific libraries. In numerical Fortran programming, current practice is to employ whatever exception handling mechanisms are provided by the system/vendor. This clearly inhibits the production of fully portable numerical libraries and programs. It is particularly frustrating now that IEEE arithmetic (specified by IEEE 754-1985 Standard for binary floating-point arithmetic, also published as IEC 559:1989, Binary floating-point arithmetic for microprocessor systems) is so widely used, since built into it are the five conditions: overflow, invalid, divide-by-zero, underflow, and inexact. Our aim is to provide support for these conditions. We have taken the opportunity to provide support for other aspects of the IEEE standard through a set of elemental functions that are applicable only to IEEE data types. This proposal involves three standard modules: IEEE_EXCEPTIONS contains a derived type, some named constants of this type, and some simple procedures.
    [Show full text]
  • Evaluating Variability Modeling Techniques for Supporting Cyber-Physical System Product Line Engineering
    This paper will be presented at System Analysis and Modeling (SAM) Conference 2016 (http://sdl-forum.org/Events/SAM2016/index.htm) Evaluating Variability Modeling Techniques for Supporting Cyber-Physical System Product Line Engineering Safdar Aqeel Safdar 1, Tao Yue1,2, Shaukat Ali1, Hong Lu1 1Simula Research Laboratory, Oslo, Norway 2 University of Oslo, Oslo, Norway {safdar, tao, shaukat, honglu}@simula.no Abstract. Modern society is increasingly dependent on Cyber-Physical Systems (CPSs) in diverse domains such as aerospace, energy and healthcare. Employing Product Line Engineering (PLE) in CPSs is cost-effective in terms of reducing production cost, and achieving high productivity of a CPS development process as well as higher quality of produced CPSs. To apply CPS PLE in practice, one needs to first select an appropriate variability modeling technique (VMT), with which variabilities of a CPS Product Line (PL) can be specified. In this paper, we proposed a set of basic and CPS-specific variation point (VP) types and modeling requirements for proposing CPS-specific VMTs. Based on the proposed set of VP types (basic and CPS-specific) and modeling requirements, we evaluated four VMTs: Feature Modeling, Cardinality Based Feature Modeling, Common Variability Language, and SimPL (a variability modeling technique dedicated to CPS PLE), with a real-world case study. Evaluation results show that none of the selected VMTs can capture all the basic and CPS-specific VP and meet all the modeling requirements. Therefore, there is a need to extend existing techniques or propose new ones to satisfy all the requirements. Keywords: Product Line Engineering, Variability Modeling, and Cyber- Physical Systems 1 Introduction Cyber-Physical Systems (CPSs) integrate computation and physical processes and their embedded computers and networks monitor and control physical processes by often relying on closed feedback loops [1, 2].
    [Show full text]
  • UML Profile for Communicating Systems a New UML Profile for the Specification and Description of Internet Communication and Signaling Protocols
    UML Profile for Communicating Systems A New UML Profile for the Specification and Description of Internet Communication and Signaling Protocols Dissertation zur Erlangung des Doktorgrades der Mathematisch-Naturwissenschaftlichen Fakultäten der Georg-August-Universität zu Göttingen vorgelegt von Constantin Werner aus Salzgitter-Bad Göttingen 2006 D7 Referent: Prof. Dr. Dieter Hogrefe Korreferent: Prof. Dr. Jens Grabowski Tag der mündlichen Prüfung: 30.10.2006 ii Abstract This thesis presents a new Unified Modeling Language 2 (UML) profile for communicating systems. It is developed for the unambiguous, executable specification and description of communication and signaling protocols for the Internet. This profile allows to analyze, simulate and validate a communication protocol specification in the UML before its implementation. This profile is driven by the experience and intelligibility of the Specification and Description Language (SDL) for telecommunication protocol engineering. However, as shown in this thesis, SDL is not optimally suited for specifying communication protocols for the Internet due to their diverse nature. Therefore, this profile features new high-level language concepts rendering the specification and description of Internet protocols more intuitively while abstracting from concrete implementation issues. Due to its support of several concrete notations, this profile is designed to work with a number of UML compliant modeling tools. In contrast to other proposals, this profile binds the informal UML semantics with many semantic variation points by defining formal constraints for the profile definition and providing a mapping specification to SDL by the Object Constraint Language. In addition, the profile incorporates extension points to enable mappings to many formal description languages including SDL. To demonstrate the usability of the profile, a case study of a concrete Internet signaling protocol is presented.
    [Show full text]
  • Mark Vie Controller Standard Block Library for Public Disclosure Contents
    GEI-100682AC Mark* VIe Controller Standard Block Library These instructions do not purport to cover all details or variations in equipment, nor to provide for every possible contingency to be met during installation, operation, and maintenance. The information is supplied for informational purposes only, and GE makes no warranty as to the accuracy of the information included herein. Changes, modifications, and/or improvements to equipment and specifications are made periodically and these changes may or may not be reflected herein. It is understood that GE may make changes, modifications, or improvements to the equipment referenced herein or to the document itself at any time. This document is intended for trained personnel familiar with the GE products referenced herein. Public – This document is approved for public disclosure. GE may have patents or pending patent applications covering subject matter in this document. The furnishing of this document does not provide any license whatsoever to any of these patents. GE provides the following document and the information included therein as is and without warranty of any kind, expressed or implied, including but not limited to any implied statutory warranty of merchantability or fitness for particular purpose. For further assistance or technical information, contact the nearest GE Sales or Service Office, or an authorized GE Sales Representative. Revised: July 2018 Issued: Sept 2005 © 2005 – 2018 General Electric Company. ___________________________________ * Indicates a trademark of General
    [Show full text]
  • Data Types and Variables
    Color profile: Generic CMYK printer profile Composite Default screen Complete Reference / Visual Basic 2005: The Complete Reference / Petrusha / 226033-5 / Chapter 2 2 Data Types and Variables his chapter will begin by examining the intrinsic data types supported by Visual Basic and relating them to their corresponding types available in the .NET Framework’s Common TType System. It will then examine the ways in which variables are declared in Visual Basic and discuss variable scope, visibility, and lifetime. The chapter will conclude with a discussion of boxing and unboxing (that is, of converting between value types and reference types). Visual Basic Data Types At the center of any development or runtime environment, including Visual Basic and the .NET Common Language Runtime (CLR), is a type system. An individual type consists of the following two items: • A set of values. • A set of rules to convert every value not in the type into a value in the type. (For these rules, see Appendix F.) Of course, every value of another type cannot always be converted to a value of the type; one of the more common rules in this case is to throw an InvalidCastException, indicating that conversion is not possible. Scalar or primitive types are types that contain a single value. Visual Basic 2005 supports two basic kinds of scalar or primitive data types: structured data types and reference data types. All data types are inherited from either of two basic types in the .NET Framework Class Library. Reference types are derived from System.Object. Structured data types are derived from the System.ValueType class, which in turn is derived from the System.Object class.
    [Show full text]
  • Adaptivfloat: a Floating-Point Based Data Type for Resilient Deep Learning Inference
    ADAPTIVFLOAT:AFLOATING-POINT BASED DATA TYPE FOR RESILIENT DEEP LEARNING INFERENCE Thierry Tambe 1 En-Yu Yang 1 Zishen Wan 1 Yuntian Deng 1 Vijay Janapa Reddi 1 Alexander Rush 2 David Brooks 1 Gu-Yeon Wei 1 ABSTRACT Conventional hardware-friendly quantization methods, such as fixed-point or integer, tend to perform poorly at very low word sizes as their shrinking dynamic ranges cannot adequately capture the wide data distributions commonly seen in sequence transduction models. We present AdaptivFloat, a floating-point inspired number representation format for deep learning that dynamically maximizes and optimally clips its available dynamic range, at a layer granularity, in order to create faithful encoding of neural network parameters. AdaptivFloat consistently produces higher inference accuracies compared to block floating-point, uniform, IEEE-like float or posit encodings at very low precision (≤ 8-bit) across a diverse set of state-of-the-art neural network topologies. And notably, AdaptivFloat is seen surpassing baseline FP32 performance by up to +0.3 in BLEU score and -0.75 in word error rate at weight bit widths that are ≤ 8-bit. Experimental results on a deep neural network (DNN) hardware accelerator, exploiting AdaptivFloat logic in its computational datapath, demonstrate per-operation energy and area that is 0.9× and 1.14×, respectively, that of equivalent bit width integer-based accelerator variants. (a) ResNet-50 Weight Histogram (b) Inception-v3 Weight Histogram 1 INTRODUCTION 105 Max Weight: 1.32 105 Max Weight: 1.27 Min Weight: -0.78 Min Weight: -1.20 4 4 Deep learning approaches have transformed representation 10 10 103 103 learning in a multitude of tasks.
    [Show full text]
  • Csci 555: Functional Programming Functional Programming in Scala Functional Data Structures
    CSci 555: Functional Programming Functional Programming in Scala Functional Data Structures H. Conrad Cunningham 6 March 2019 Contents 3 Functional Data Structures 2 3.1 Introduction . .2 3.2 A List algebraic data type . .2 3.2.1 Algebraic data types . .2 3.2.2 ADT confusion . .3 3.2.3 Using a Scala trait . .3 3.2.3.1 Aside on Haskell . .4 3.2.4 Polymorphism . .4 3.2.5 Variance . .5 3.2.5.1 Covariance and contravariance . .5 3.2.5.2 Function subtypes . .6 3.2.6 Defining functions in companion object . .7 3.2.7 Function to sum a list . .7 3.2.8 Function to multiply a list . .9 3.2.9 Function to remove adjacent duplicates . .9 3.2.10 Variadic function apply ................... 11 3.3 Data sharing . 11 3.3.1 Function to take tail of list . 12 3.3.2 Function to drop from beginning of list . 13 3.3.3 Function to append lists . 14 3.4 Other list functions . 15 3.4.1 Tail recursive function reverse . 15 3.4.2 Higher-order function dropWhile . 17 3.4.3 Curried function dropWhile . 17 3.5 Generalizing to Higher Order Functions . 18 3.5.1 Fold Right . 18 3.5.2 Fold Left . 22 3.5.3 Map . 23 1 3.5.4 Filter . 25 3.5.5 Flat Map . 26 3.6 Classic algorithms on lists . 27 3.6.1 Insertion sort and bounded generics . 27 3.6.2 Merge sort . 29 3.7 Lists in the Scala standard library .
    [Show full text]
  • Handwritten Digit Classication Using 8-Bit Floating Point Based Convolutional Neural Networks
    Downloaded from orbit.dtu.dk on: Apr 10, 2018 Handwritten Digit Classication using 8-bit Floating Point based Convolutional Neural Networks Gallus, Michal; Nannarelli, Alberto Publication date: 2018 Document Version Publisher's PDF, also known as Version of record Link back to DTU Orbit Citation (APA): Gallus, M., & Nannarelli, A. (2018). Handwritten Digit Classication using 8-bit Floating Point based Convolutional Neural Networks. DTU Compute. (DTU Compute Technical Report-2018, Vol. 01). General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Handwritten Digit Classification using 8-bit Floating Point based Convolutional Neural Networks Michal Gallus and Alberto Nannarelli (supervisor) Danmarks Tekniske Universitet Lyngby, Denmark [email protected] Abstract—Training of deep neural networks is often con- In order to address this problem, this paper proposes usage strained by the available memory and computational power. of 8-bit floating point instead of single precision floating point This often causes it to run for weeks even when the underlying which allows to save 75% space for all trainable parameters, platform is employed with multiple GPUs.
    [Show full text]